Calculated Results of $(n,2n)$ and $(n,3n)$ Reaction Excitation Functions Using a Combined Systematics and Neural Network Method

The original data were subjected to outlier screening, consistency checks, and training/test splitting before being used to train the NN+SYS model. The trained model was then used to calculate the reaction cross sections of each nuclide as a function of incident neutron energy.The data released in this dataset include only the NN+SYS predictions for the nuclides investigated in the study. Each row in the data table represents an NN+SYS prediction for a specific target nuclide at a given incident neutron energy.

The main fields include the proton number Z, mass number A, incident neutron energy E, and the reaction cross section predicted by NN+SYS. Both Z and A are dimensionless integers. The incident neutron energy E is given in MeV, and the predicted reaction cross section is given in barns (b).

Multiple records are provided for the same nuclide at different incident neutron energies, and these data points collectively constitute the corresponding (n,2n) or (n,3n) reaction excitation function.